{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/tapex-table-pre-training-via-learning-a","title":"TAPEX: Table Pre-training via Learning a Neural SQL Executor","arxiv_id":"2107.07653","date":"2021-07-16","proceeding":"ICLR 2022 4","authors":["Qian Liu","Bei Chen","Jiaqi Guo","Morteza Ziyadi","Zeqi Lin","Weizhu Chen","Jian-Guang Lou"],"abstract":"Recent progress in language model pre-training has achieved a great success via leveraging large-scale unstructured textual data. However, it is still a challenge to apply pre-training on structured tabular data due to the absence of large-scale high-quality tabular data. In this paper, we propose TAPEX to show that table pre-training can be achieved by learning a neural SQL executor over a synthetic corpus, which is obtained by automatically synthesizing executable SQL queries and their execution outputs. TAPEX addresses the data scarcity challenge via guiding the language model to mimic a SQL executor on the diverse, large-scale and high-quality synthetic corpus. We evaluate TAPEX on four benchmark datasets. Experimental results demonstrate that TAPEX outperforms previous table pre-training approaches by a large margin and achieves new state-of-the-art results on all of them. This includes the improvements on the weakly-supervised WikiSQL denotation accuracy to 89.5% (+2.3%), the WikiTableQuestions denotation accuracy to 57.5% (+4.8%), the SQA denotation accuracy to 74.5% (+3.5%), and the TabFact accuracy to 84.2% (+3.2%). To our knowledge, this is the first work to exploit table pre-training via synthetic executable programs and to achieve new state-of-the-art results on various downstream tasks. Our code can be found at https://github.com/microsoft/Table-Pretraining.","url_abs":"https://arxiv.org/abs/2107.07653v3","url_pdf":"https://arxiv.org/pdf/2107.07653v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"tapex-table-pre-training-via-learning-a","repo_url":"https://github.com/microsoft/Table-Pretraining","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"tapex-table-pre-training-via-learning-a","repo_url":"https://github.com/sohanpatnaik106/cabinet_qa","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"tapex-table-pre-training-via-learning-a","repo_url":"https://github.com/MindCode-4/code-5/tree/main/tapex","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"tapex-table-pre-training-via-learning-a","repo_url":"https://github.com/pwc-1/Paper-9/tree/main/1/tapex","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"semantic-parsing","task_name":"Semantic Parsing"},{"task_slug":"table-based-fact-verification","task_name":"Table-based Fact Verification"}],"methods":[{"method_slug":"tapex","method_name":"TAPEX"}],"datasets_introduced":[],"methods_introduced":[{"slug":"tapex","name":"TAPEX","full_name":"Table Pre-training via Execution"}],"results":[{"leaderboard":"/sota/semantic-parsing-on-sqa","task":"Semantic Parsing","dataset":"SQA","model":"TAPEX-Large","rank_in_archive_order":1,"of":2,"metrics":{"Denotation Accuracy":"74.5"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-parsing-on-wikisql-1","task":"Semantic Parsing","dataset":"WikiSQL","model":"TAPEX-Large (weak supervision)","rank_in_archive_order":2,"of":5,"metrics":{"Denotation accuracy (test)":"89.5"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-parsing-on-wikitablequestions","task":"Semantic Parsing","dataset":"WikiTableQuestions","model":"TAPEX-Large","rank_in_archive_order":17,"of":22,"metrics":{"Accuracy (Dev)":"57.0","Accuracy (Test)":"57.5"},"uses_additional_data":false},{"leaderboard":"/sota/table-based-fact-verification-on-tabfact","task":"Table-based Fact Verification","dataset":"TabFact","model":"TAPEX-Large","rank_in_archive_order":8,"of":15,"metrics":{"Test":"84.2","Val":"84.6"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2107.07653","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}